Available data indicate that a significant proportion of homosexual men continue to engage in sexual behavior that places them at high risk for AIDS, despite awareness of "safe sex" recommendations. Existing models of preventive health behavior are of limited value in explaining the persistenance of such behavior because they omit important factors specific to the threat of AIDS and the nature of the behavior to be altered. The proposed study offers a new, provisional model of the factors that influence gay men's sexual behavior in the context of the AIDS epidemic. In addition to traditional variables, such as perceived risk of getting AIDS, perceived efficacy of health recommendations, and costs and benefits of action, we have added centrality of gay sex to one's gay identify and lifestyle, perceived credibility of the source(s) of the recommendations, peer group norms, partner pressure, personal contact with individuals with AIDS, and past successes or failures of one's previous attempts at change. These factors emerged from our own past research, our clinical practice with this population and the work of other investigators in the field. The sample consists of 210 research subjects; one-third will be gay men who have tested positive for the HIV antibody, one-third will be gay men who have tested negative, and the remainder will not have undergone testing. Sixty of these men (20 from each group) will participate in small group interviews. Groups will be homogeneous with respect to antibody status (i.e. positive or unknown). One hundred-fifty men (50 in each group) will be individually interviewed. The research design is innovative; it intergrates qualitative data gathering techniques in the form of group and individual interviews, with quantitative data analysis. Unstructured interviewing techniques will be used instead of structured interviews with standardized measures. This has the advantage of allowing the respondent to "define the situation" from his own perspective, not the researchers'. It also assures that all factors important in determining sexual behavior will have the opportunity to emerge, even those not anticipated at the study outset. The results will provide an in-depth understanding of gay men's sexual decision-making within the context of the threat of AIDS. Such an understanding would greatly enhance our ability to develop effective public health interventions targeted to the at- risk gay male population.